Proc. Natl. Acad. Sci. USA Vol. 90, pp. 2710-2714, April 1993 Neurobiology Adenine nucleoside diphosphates block adaptation of mechanoelectrical transduction in hair cells (auditory system/ear/molecular motor/myosin/vestibular system) PETER G. GILLESPIE AND A. J. HUDSPETH Center for Basic Neuroscience Research, University of Texas Southwestern Medical Center, Dallas, TX 75235-9039 Contributed by A. J. Hudspeth, December 16, 1992

ABSTRACT By adapting to sustained stimuli, hair cells in the internal ear retain their sensitivity to minute transient ATP ATP displacements. Because one model for adaptation asserts that this process is mediated by a myosin isozyme, we reasoned that we should be able to arrest adaptation by interfering with ATPase cycle though introduction of ADP into hair myosin's ADPj I, Weak- cells. During tight-seal, whole-cell recordings of transduction binding currents in cells isolated from bullfrog (Rana catesbeiana) states sacculus, dialysis with 5-25 mM ADP gave variable results. In half of the cells examined, the rate of adaptation remained unchanged or even increased; adaptation was blocked in the ADP remaining cells. Because we suspected that the variable effect of ADP resulted from the conversion of ADP to ATP by PTrl adenylate kinase, we employed the ADP analog adenosine 5'-[(3-thio]diphosphate (ADP[.JS]), which is not a substrate for adenylate kinase. Adaptation consistently disappeared in the Power stroke presence of 1-10 mM ADP[(ISJ; in addition, the transduction channels' open probability at rest grew from -0.1 to 0.8 or FIG. 1. Schematic depiction of myosin's ATPase cycle. In the more. Both effects could be reversed by 2 mM ATP. When used nucleotide-free or rigor state, the myosin molecule is tightly bound to an actin filament. ATP binds to myosin and promotes its rapid in conjunction with the adenylate kinase inhibitorP1,P5-bis(5'- dissociation from the filament; during the ensuing weak interaction adenosyl) pentaphosphate (Ap5A), ADP had effects similar to with actin, myosin cleaves ATP to ADPPi. As Pi is released, the those ofADP[(8S]. These results suggest that adaptation by hair myosin molecule rebinds tightly to the actin filament and performs a cells involves adenine nucleotides, and they lend support to the power stroke. Finally, ADP dissociates and the rigor state recurs. hypothesis that the adaptation process is powered by a myosin Because myosin and actin interact with high affinity, the rigor and motor. ADP-bound states are both capable of sustaining mechanical force.

By responding to sounds and accelerations, the internal ear ity is regulated by the cytoplasmic Ca2+ concentration (12, performs one type of mechanoelectrical transduction. Hair 16, 18; reviewed in ref. 2). During adaptation to positive cells, the receptors of the cochlea and vestibular labyrinth, stimuli, it is supposed that Ca2+ enters a stereocilium through transform mechanical stimuli into electrical signals (reviewed open transduction channels and causes the motor to slip in refs. 1 and 2). The electrical response of each such cell is down a microfilamentous track, relieving tension in the gating produced by deflection ofa mechanically sensitive organelle, spring. During negative stimuli, when transduction channels the hair bundle, which emerges from the hair cell's apical are closed and the cytoplasmic Ca2+ concentration falls, the surface. A hair bundle comprises 20-300 actin-stiffened motor climbs up the track to restore tension. Because it could stereocilia, clustered together like the staggered pipes of an move along the microfilaments that form the cytoskeletal organ, and a single true cilium, the kinocilium. When a hair core of a stereocilium (19), a member of the myosin family bundle is displaced, transduction channels, perhaps as few as might well serve as the adaptation motor (12, 18, 20). one per stereocilium (3-5), are directly pulled open by To test the hypothesis that myosin molecules are respon- tension in gating springs within the hair bundle (6). Because sible for adaptation in the hair cell's transduction process, we these channels occur at the bundle's top (7, 8), it is probable sought to arrest the motors upon their tracks. A hair cell with that each gating spring is the thin, filamentous tip link that its adaptation motors stalled would be incapable of adjusting extends from the end ofa stereocilium to the side ofits tallest tip-link tension and should therefore maintain a constant neighbor (9, 10). transduction current in response to a protracted bundle When a hair bundle is subjected to a protracted displace- our approach to the ATPase cycle of ment, the cell's response adapts (11-16): the bundle's range displacement. Tailoring of responsiveness migrates toward the position to which the skeletal-muscle myosin (Fig. 1; reviewed in refs. 21 and 22), bundle has been moved. Adaptation is associated with me- we sought to arrest a myosin molecule in either of two states chanical changes within the hair bundle that are thought to in the cycle: rigor, in which the myosin molecule has no reflect adjustment of the tension in gating springs (12, 17). bound nucleotide, or the ADP-bound state, which ensues One model for adaptation posits that the tension in each after release of inorganic phosphate ion (Pj). A myosin gating spring is maintained by motor molecules whose activ- molecule confined to either of these states would be tightly bound to the actin fiament, able neither to climb nor to slip. The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" Abbreviations: ADP[,8S], adenosine 5'-[f-thio]diphosphate; ApsA, in accordance with 18 U.S.C. §1734 solely to indicate this fact. P1,P5-bis(5'-adenosyl) pentaphosphate. 2710 Downloaded by guest on September 23, 2021 Neurobiology: Gillespie and Hudspeth Proc. Natl. Acad. Sci. USA 90 (1993) 2711 Our results, some of which have appeared in preliminary I 1 form (23), demonstrate that ADP or an ADP analog blocks adaptation, and thus implicate adenine nucleotides in adap- Imax 1 + e&z(X-Xo)/kT' tation. Furthermore, the characteristics of the inhibition of adaptation are entirely consistent with the suggestion that the Here p is the channels' open probability, I is the transduction process is myosin based. current in response to a hair-bundle displacement X, Ima is the maximal transduction current, z is the channels' sensi- tivity to displacement, and Xo is the displacement at which MATERIALS AND METHODS half the channels are open; k is the Boltzmann constant and Adenine Nucleotide Solutions. Adenine nucleotides were T is the thermodynamic temperature. We used the fitted data obtained from Boehringer Mannheim. The concentrations of to estimate the fraction ofchannels open with the hair bundle ATP, ADP, and adenosine 5'-[,-thio]diphosphate (ADP[/S3]) in its resting position. To demonstrate that changes in the were determined by their absorbances at 259 nm on the basis channels' open probability were not artifactually produced of a molar extinction coefficient of 1.54 x 106 M-1m-1 (24). by drift of the stimulus probe, recording electrode, or micro- We estimated the concentration of Pl,P5-bis(5'-adenosyl) scope stage, we occasionally detached the probe and re- pentaphosphate (Ap5A) by assuming twice that molar extinc- attached it after confirming that the bundle was in its resting tion coefficient. The purity ofthe nucleotides was established position. by high-performance liquid chromatography on an anion- Transduction currents were measured with an amplifier exchange column (Mono Q, Pharmacia/LKB) with a 15-ml (EPC-7, List Electronics, Darmstadt, Germany) operated gradient of 0-1 M ammonium bicarbonate. Although the without series-resistance compensation. Stimulation was purities of ATP and ADP exceeded 95%, ADP[f3S] was controlled and responses were recorded by means of an contaminated by AMP (15%) and a small amount of ADP experimental interface (Indec Systems, Sunnyvale, CA) (0.1%). whose operation was programmed in BASIC-23 on a computer Electrophysiological Recording. Hair cells were isolated (PDP-11/73, Digital Equipment). The experimental environ- from bullfrog (Rana catesbeiana) sacculus after exposure of ment has been described in detail (8). the organ in situ to low-Ca2+ saline solution (10, 16). In a modification of the published procedure, 1 mM EGTA and 1 mM MgCl2 replaced 1 mM EDTA in the solution dripped into RESULTS the perilymphatic cistern. In addition, DNase treatment was We used whole-cell, tight-seal electrodes to record transduc- omitted and cells were mechanically dissociated in saline tion currents from hair cells isolated from the bullfrog's solution containing 100 iM CaC12. sacculus. When a pipette contained no nucleotide or milli- After solitary hair cells were firmly adsorbed onto a molar concentrations ofATP, our results accorded with those concanavalin A-coated coverslip in an experimental cham- from previous investigations of hair cells from this and other ber, they were maintained at 21°C in an oxygenated saline organs (reviewed in ref. 1). With the hair bundle in its resting solution containing 110 mM Na+, 2 mM K+, 4 mM Ca2+, 118 position, the cell's mechanically sensitive channels bore an mM Cl-, 3 mM D-glucose, and 5 mM Hepes at pH 7.25. Cells inward transduction current that was -15% of the maximum were observed through a 63x objective lens and Nomarski that could be evoked; the channels' open probability was thus differential-interference-contrast optics on an inverted mi- -0.15 (3). Moving the hair bundle progressively in the croscope (Zeiss model IM-35) with a rotating stage. positive direction, towards its tall edge, opened additional Transduction currents were recorded with whole-cell tight- channels until it evoked a transduction current that could be seal pipettes (25), whose axial resistances were 3-4 MfQ prior as great as -375 pA. Negative stimuli, on the other hand, to seal formation and generally rose to 10-20 MQI during closed some or all ofthe channels that were open at rest (Fig. recording. All internal pipette solutions contained 85 mM 2A). CsCl, 1 mM MgCl2, 1 mM EGTA, and 5 mM Hepes. After During a given displacement, the transduction current did addition of adenine nucleotide with an equimolar concentra- not persist at a constant amplitude (11, 13, 14, 16). Instead, tion of MgCl2, an internal solution's pH was adjusted to as adaptation occurred in response to a positive displace- 7.25-7.30 with CsOH. Recordings were made at a holding ment, the response declined towards the resting level. Ad- potential of -74 mV, which took into account dissipation of aptation also proceeded during a negative displacement; the a +4-mV tip potential between a pipette's contents and the amplitude of the current transient following such a stimulus bath solution. Because ADP[f3B] was supplied as its lithium revealed the degree of adaptation. salt, we performed several control experiments with 10 mM When the recording pipette contained millimolar concen- LiCl or 20 mM LiCl plus 2 mM ATP added to the internal trations ofATP, adaptation was robust (Fig. 2A) and endured solution; transduction and adaptation in these instances were for as long as the cell was healthy, generally for more than 15 similar to those under ordinary conditions. min. When the pipette contained no nucleotide, a cell's In some experiments, we filled the tip of the recording ability to adapt was labile, but nevertheless often persisted electrode with a solution different from that in the pipette's for several minutes. If ATP is required for the activity of an shank. To ensure that the shank solution was capable of adaptation motor, the cells from which we recorded evidently diffusing into a cell in only a few minutes, we dipped the contained a substantial reservoir ofthe compound or retained electrode's tip into the appropriate solution for only 1-2 s. the capacity to produce more. Control experiments with dye-containing solutions indicated Because we anticipated difficulty in extensively lowering that this procedure provided -2 nl of fluid at the pipette's tip. the cellular ATP concentration through a micropipette (26), Stimuli were delivered with a micropipette, which was we attempted instead to arrest the putative hair-bundle attached to the kinocilium's bulbous tip by gentle suction (3) motors by dialyzing hair cells with ADP. In half (6 of 12) of and displaced by a piezoelectrical stack (PZL-007-1, Burleigh the hair cells dialyzed with 5-25 mM ADP, adaptation was, Instruments, Fishers, NY). We monitored the time course as expected, blocked within a few minutes (data not shown). and extent of adaptation to 100-ms bundle displacements of In the remaining 6 cells, by contrast, adaptation persisted or up to +500 nm. For the construction of displacement- even accelerated; several cells retained robust transduction response relations, cells were subjected to a range of 16-ms and adaptation for longer than 20 min. Because such enduring deflections of amplitudes up to ±1050 nm. The peak trans- adaptation was reminiscent of the finding in control cells duction currents were fitted with a Boltzmann relation (4, 6), dialyzed with ATP, we suspected that adenylate kinase, the Downloaded by guest on September 23, 2021 2712 Neurobiology: Gillespie and Hudspeth Proc. Natl. Acad. Sci. USA 90 (1993) A ATP B ADPIS C ADP/Ap5A c - _ 100 s

oi- 100A 10PA 100 pA

c 840 s

p4--'I 600 s oj- 42 s L | | ~~~~~~~~~~~~~~~11500 nm 50 ms FIG. 2. Arrest ofadaptation by ADP[,BS] or by ADP and ApNA. Families oftransduction-current records were elicited by 100-ms hair-bundle displacements of the magnitudes shown in the bottom traces. In this and subsequent figures, the time indicated beside a record represents the interval from the onset of whole-cell recording until collection ofthe displayed data commenced. For each set of traces, a pair offiducial marks indicates the current with all the channels closed (labeled c in part A) or open (labeled o). (A) During recording with a pipette filled with solution containing 2 mM ATP, brisk adaptation persisted for more than 14 min. Ten traces were averaged for the set ofresponses obtained at each time. (B) When the recording pipette contained 10 mM ADP[,BS], adaptation vanished as the nucleotide diffused into a cell. During the same period, the probability that the transduction channels were open at rest increased from 0.17 to 0.80. Because ofthe rapidity of the ADP[/3S] effect, only three traces were averaged at each time. (C) A similar effect was observed when the recording pipette contained 5 mM ADP and 250 ,M ApsA. In this instance, the recording was begun with 2 mM ATP in the pipette's tip, and adaptation was gradually arrested as that nucleotide was replaced by diffusion of ADP and Ap5A from the pipette's shank. Each record represents the average of 10 traces. ubiquitous enzyme that catalyzes the interconversion of with the bundle in its resting position increased from -0.1 to adenine nucleotides (27), continuously produced sufficient 0.8 or more. The displacement-response relation, which ATP in ADP-filled hair cells to support adaptation. characterizes the probability that channels are opened by To preclude adenylate kinase activity, we dialyzed hair stimuli of various amplitudes, consequently shifted in the cells with ADP[/3S], an ADP analog that is not a substrate for negative direction (Fig. 4). After the channels' open proba- the enzyme (28). In 26 of29 cells filled with 10 mM ADP[,fS], bility at rest exceeded 0.8, it sometimes fell spontaneously to adaptation was dramatically reduced during both positive and 0.5 or less. The latter phenomenon occurred more frequently negative bundle displacements (Fig. 2B). The loss of adap- in cells dialyzed with 0.1-1 mM ADP[j3S] than in those tation usually occurred promptly; in most cells, adaptation exposed to a higher concentration of the nucleotide. vanished within 60 s of establishment of the whole-cell In many cells dialyzed with ADP[,8S], transduction sud- recording configuration (Fig. 3A). Adaptation was also ar- denly disappeared several minutes after the arrest of adap- rested within 120 s in four of five cells dialyzed with 1 mM tation. In several other instances, the conventional response ADP[f3S]. For each of the four cells that did not lose abruptly vanished, but a transient current, presumably borne adaptation when filled with 1-10 mM ADP[,BS], the recording by transduction channels, could still be elicited (Fig. 3 B pipette's access resistance was unusually high. When six and C). cells were filled with 100 ,.M analog, two cells showed no We demonstrated in two cells that the effects of ADP[,3S] effect, and blockage required 200-500 s in the remainder. At could be almost entirely reversed by ATP. When a pipette's a concentration of 1 mM or greater, ADP[,BS] thus appeared tip was filled with 10 mM ADP[,BS] and its shank with 2 mM to stall the adaptation motors in the manner expected. ATP, adaptation was blocked shortly after the onset of As it arrested adaptation, ADP[j3S] treatment consistently recording. As ATP subsequently diffused into the cell, how- caused an additional change in a hair cell's mechanosensi- ever, adaptation resumed and the resting open probability tivity: the transduction channels' probability of being open returned to approximately its control value (Fig. 5). A

gf 23 s gM 26 s 29 s 32 s 1100 pA - FIG. 3. Time course of the effects of ADP[/3S] 100 ms on transduction and adaptation by a single cell. (A) At each of the indicated times, families of trans- 35 s 38 s 41 s 44 s duction currents were measured with the stimulus paradigm displayed in Fig. 2B. The reduction in the rate of adaptation and the increase in the channels' resting open probability developed continuously E3 C and rapidly. Individual records are displayed. (B) After acquisition of the last record shown in A, 'T were Wf7 1100 r 100 pA transduction and adaptation abruptly and al- pA_ most entirely lost; thereafter, transient inward cur- - rents were sporadically elicited by positive stimuli. s 100 ms (C) Displayed on a faster time base, a portion ofthe record shown in B (indicated by an arrow) revealed J---L- 1500 nm 1500 nm 1500 nm the transient nature of the residual current. Downloaded by guest on September 23, 2021 Neurobiology: Gillespie and Hudspeth Proc. Natl. Acad. Sci. USA 90 (1993) 2713

A ATP 1 0 s t 1.0 - c- n l r 0.8 - 84 s Cu n 0.6 - 0- 0 C- SOs a10.Q0.4 - oQ 0.2 [/ 780 s 0.0 0- __ 1100 pA 240 s -1000 -500 0 500 1000 C - B 100 ms 1.0 0- = 0.8 -T~- 1500 nm Cuo 0.6 S.0 XL 0.4 FIG. 5. Reversal by ATP of the effects of ADP[,8S] on transduc- a' 0.2 0 tion and adaptation. Transduction currents were recorded with a 0.0 pipette whose tip was filled with 10 mM ADP[,BS] and shank with 2 mM ATP. Ten seconds after establishment of the whole-cell config- -1000 -500 0 500 1000 uration, adaptation was normal. After 50 s of recording, adaptation was mostly blocked and the channels' resting open probability was Ca ADP/Ap5A 0.92. Later in the recording session (240 s), the effects of ADP[f3S] 1.0 were largely reversed as ATP diffused into the cell from the pipette's = 460 s shank. Three sets of transduction currents were averaged in each 0.8 - case. During the final record displayed, and still more strikingly at times between the 50-s and 240-s records, the channels' open co probability systematically increased after successive positive stim- uli. The open probability then reverted to a lower value after a large 0.2- _/\130 s 0 negative stimulus. 0.0 -. .~~~ esis that adaptation is effected by a myosin isozyme. Adap- -1000 -500 0 500 1000 tation was rapidly and completely blocked as ADP[13S] Displacement (nm) diffused into a cell; in addition, ADP often had a similar FIG. 4. Displacement-response relations obtained during dialy- effect. Because mechanoelectrical transduction itself re- sis of hair cells with 2 mM ATP (A), 10 mM ADP[3S] (B), or 5 mM mained robust, it appeared as if the adaptation mechanism ADP and 250 ,uM Ap5A (with the pipette's tip initially filled with 2 had been selectively arrested. The transduction channels' mM ATP) (C). The displacement-response relation shifted in the probability of being open at rest increased, which likely negative direction and became shallower in the presence ofADP[j3S]; reflected greater tension applied to the gating springs by both effects were much less pronounced with the mixture ofADP and adaptation motors. This phenomenon is consistent with a Ap5A. Fits ofthe data shown in B with a Boltzmann relation indicated feature of conventional myosin isozymes: in isometrically that Xo, the point at which half of the channels were open, shifted contracting skeletal muscle provided with ATP, tension from +120 nm to -381 nm as ADP[,BS] filled the cell. In addition, z, the single-channel gating force and a measure of the curve's slope, increases by as much as 50% upon exposure to millimolar declined from 89 to 20 fN. This decreased slope may have reflected concentrations of ADP (30, 31). Because myosin molecules dispersion ofthe sensitive ranges ofindividual transduction elements sustain force in tightly bound states, this effect is thought to (1). The maximal transduction current was -102 pA at the record- reflect an augmented population ofmyosin molecules trapped ing's outset and -98 pA at its conclusion. in force-producing states after their power strokes (31). Because the transduction channels' probability ofbeing open Another potential means of avoiding the effects of adenyl- at rest depends on the tension in gating springs (10, 17), an ate kinase is to inhibit the enzyme. Some isozymes of increase in the population ofthe ADP-bound state well adenylate kinase are inhibited In might potently by Ap5A (29). each raise the open In of the seven cells that we successfully dialyzed with a probability. the cell whose mechanical sensitivity was characterized by the Boltzmann relation combination of 5 mM ADP and 250 ,uM ApNA, adaptation slowed over 120-600 s (Fig. 2C). Although with this treat- presented in Fig. 4B, it may be calculated (32) that a 50o ment the open probability at rest again rose, its value rarely increase in gating-spring tension would have raised the chan- exceeded 0.5 (Fig. 4C). An additional, unexplained, effect nels' open probability at rest from -0.1 to >0.9. was observed: adaptation was usually absent immediately If the increase in open probability resulted from enhanced after the whole-cell recording configuration was established tension in gating springs, it should have been accompanied by with a pipette containing ADP and Ap5A but then recovered movement of an unconstrained hair bundle in the negative over a minute or so. This transient effect was evidently due direction (16). The expected motion (32), however, would to extracellular leakage of Ap5A as the pipette approached have been only some -70 nm. Because of the drift inherent the cell. To avoid this complication, we made recordings in in present displacement-measuring systems (12, 16, 17), a which a pipette with similar contents was filled at its tip with movement of this magnitude over a period of many seconds a few nanoliters of 2 mM ATP. In each of four experiments, cannot yet be detected. the initial interruption of adaptation was then absent, and the After protracted dialysis of a cell with ADP[PS], ordinary subsequent decline of adaptation resembled that observed mechanical responses sometimes vanished and the resting with ADP[,3S] in the pipette solution (Fig. 2C). open probability approached zero (Fig. 3B). This phenome- non resembles that observed upon exposure of hair bundles to an extracellular Ca2+ concentration below -20 ,uM (5), DISCUSSION which results in the destruction of tip links (10), the pre- The interruption of adaptation by adenine nucleoside diphos- sumptive gating springs. Because ADP[13S] treatment some- phates displayed characteristics consistent with the hypoth- times spared intermittent and transient responsiveness to Downloaded by guest on September 23, 2021 2714 Neurobiology: Gillespie and Hudspeth Proc. NatL Acad Sci. USA 90 (1993) mechanical stimulation (Fig. 3C), however, the nucleotide 1. Howard, J., Roberts, W. M. & Hudspeth, A. J. (1988) Annu. was unlikely to have destroyed the tip links. What, then, Rev. Biophys. Biophys. Chem. 17, 99-124. might have been the source of the residual responses? It is 2. Hudspeth, A. J. (1989) Nature (London) 341, 397-404. plausible that the motor associated with each tip link contains 3. Holton, T. & Hudspeth, A. J. (1986) J. Physiol. (London) 375, several myosin molecules (17, 32). After ADP[,/S] had inac- 195-227. tivated many of these, enhanced tension in the tip links may 4. Howard, J. & Hudspeth, A. J. (1988) Neuron 1, 189-199. 5. Crawford, A. C., Evans, M. G. & Fettiplace, R. (1991) J. well have torn the motors loose from their cytoskeletal Physiol. (London) 434, 369-398. attachments. Transient currents might then have arisen 6. Corey, D. P. & Hudspeth, A. J. (1983) J. Neurosci. 3, 962-976. whenever evanescent rebinding of a few myosin molecules 7. Hudspeth, A. J. (1982) J. Neurosci. 2, 1-10. allowed the gating springs to sustain some tension. 8. Jaramillo, F. & Hudspeth, A. J. (1991) Neuron 7, 409-420. The combination of ADP and Ap5A inhibited adaptation 9. Pickles, J. O., Comis, S. D. & Osborne, M. P. (1984) Hearing considerably more slowly than did ADP[f8S]. It is possible Res. 15, 103-112. that the concentration of Ap5A employed, although substan- 10. Assad, J. A., Shepherd, G. M. G. & Corey, D. P. (1991) Neu- tial, was nonetheless insufficient to fully inhibit the hair cell's ron 7, 985-994. adenylate kinase. Soluble adenylate kinases are blocked by 11. Eatock, R. A., Corey, D. P. & Hudspeth, A. J. (1987) J. Ap5A with inhibition constants (K, values) of 2-20 nM with Neurosci. 7, 2821-2836. respect to ADP (29, 33, 34), but mitochondrial isozymes are 12. Howard, J. & Hudspeth, A. J. (1987) Proc. Natl. Acad. Sci. much less sensitive to the compound, with con- USA 84, 3064-3068. inhibition 13. Assad, J. A., Hacohen, N. & Corey, D. P. (1989) Proc. Natl. stants of 2-6 ,uM (34, 35). If Ap5A is a competitive inhibitor Acad. Sci. USA 86, 2918-2922. with respect to ADP (35), the 20-fold higher concentration of 14. Crawford, A. C., Evans, M. G. & Fettiplace, R. (1989) J. ADP in our internal solution may have promoted the syn- Physiol. (London) 419, 405-434. thesis of enough ATP to support adaptation. 15. Hacohen, N., Assad, J. A., Smith, W. J. & Corey, D. P. (1989) In the light of the arguments presented above, it seems J. Neurosci. 9, 3988-3997. peculiar that hair cells can retain adaptation for minutes when 16. Assad, J. A. & Corey, D. P. (1992) J. Neurosci. 12, 3291-3309. dialyzed with internal solutions lacking adenine nucleotides 17. Jaramillo, F. & Hudspeth, A. J. (1993) Proc. Natl. Acad. Sci. altogether (13, 14). If the initial cytoplasmic concentration of USA 90, 1330-1334. ATP was =1 mM (36), however, dialysis ofa hair cell through 18. Howard, J. & Hudspeth, A. J. (1987) in Sensory Transduction, pipettes such as those we employed should have left the ATP eds. Hudspeth, A. J., MacLeish, P. R., Margolis, F. L. & concentration above -100 ,uM for at least 10 min (26). This Wiesel, T. N. (Fond. Etude Syst. Nerv. Cen. Periph., Geneva), pp. 138-145. concentration is sufficient to sustain many ATP-dependent 19. Shepherd, G. M. G., Corey, D. P. & Block, S. M. (1990) Proc. processes, including myosin-based motility (30). Natl. Acad. Sci. USA 87, 8627-8631. It is appropriate to inquire whether other possible mech- 20. Gillespie, P. G. & Hudspeth, A. J. (1992) Soc. Neurosci. Abstr. anisms of adaptation, besides that invoking myosin, could 18, 1398. explain our results. Because ADP[,BS] may inhibit adenylate 21. Hibberd, M. G. & Trentham, D. R. (1986) Annu. Rev. Biophys. kinase (28), any cellular process dependent upon the en- Biophys. Chem. 15, 119-161. zyme's activity might have been blocked. A reduced intra- 22. Cooke, R. (1986) CRC Crit. Rev. Biochem. 21, 53-118. cellular Ca2+ concentration, such as that produced by depo- 23. Gillespie, P. G. & Hudspeth, A. J. (1992) Biophys. J. 61, 516 larization or by lowering the extracellular Ca2+ concentra- (abstr.). 24. Fasman, G. D., ed. (1975) Handbook of Biochemistry and tion, also diminishes the rate of adaptation and shifts the Molecular Biology, Nucleic Acids (CRC, Cleveland), Vol. 1, displacement-response relationship (13, 14, 16). Because the 3rd Ed., p. 149. amplitude of the transduction current was usually unaffected 25. Hamill, 0. P., Marty, A., Neher, E., Sakmann, B. & Sigworth, by ADP[f3S] or by ADP and ApNA, however, it is unlikely that F. J. (1981) Pflugers Arch. 391, 85-100. the compounds blocked the only known route of Ca2+ entry 26. Pusch, M. & Neher, E. (1988) Pfldgers Arch. 411, 204-211. into the hair bundle, that through the transduction channels 27. Noda, L. (1973) in The Enzymes, ed. Boyer, P. D. (Academic, (37, 38). It is also improbable that ADP[j8S], ADP, or Ap5A New York), Vol. 8, Part A, 3rd Ed., pp. 279-305. accelerated Ca2+ extrusion by a pump or exchanger. In 28. Goody, R. S., Eckstein, F. & Schirmer, R. H. (1972) Biochim. particular, the ATP requirement ofCa2+-ATPase implies that Biophys. Acta 276, 155-161. 29. Lienhard, G. E. & Secemski, I. I. (1973) J. Biol. Chem. 248, the nucleotides would more likely have inhibited Ca2+ re- 1121-1123. moval by a pump. The effects of the nucleoside diphosphates 30. Cooke, R. & Pate, E. (1985) Biophys. J. 48, 789-798. are also difficult to reconcile with an alternative model of 31. Pate, E. & Cooke, R. (1989) J. Muscle Res. Cell Motil. 10, adaptation that involves stabilization of a closed state of the 181-1%. channel by Ca2+ (14). Finally, microtubule-based motor 32. Hudspeth, A. J. (1992) in Sensory Transduction, eds. Corey, proteins such as kinesin and dynein cannot be responsible for D. P. & Roper, S. D. (Rockefeller Univ. Press, New York), pp. adaptation; the only microtubules found in the hair bundle 357-370. occur in the kinocilium, which is unnecessary both for 33. Feldhaus, P., Frohlich, T., Goody, R. S., Isakov, M. & transduction (39) and for adaptation (16). Although adapta- Schirmer, R. H. (1975) Eur. J. Biochem. 57, 197-204. 34. Kuby, S. A., Hamada, M., Gerber, D., Tsai, W.-C., Jacobs, tion might be effected by an unknown process that exhibits H. K., Cress, M. C., Chua, G. K., Fleming, G., Wu, L. H., the characteristics reported here, the hypothesis that a my- Fischer, A. H., Frischat, A. & Maland, L. (1978) Arch. Bio- osin isozyme mediates adaptation most parsimoniously fits chem. Biophys. 187, 34-52. our data. 35. Font, B. & Gautheron, D. C. (1980) Biochim. Biophys. Acta 611, 299-308. We relied upon the expert technical support of Mr. R. A. Jacobs, 36. Chapman, A. G., Fall, L. & Atkinson, D. E. (1971) J. Bacte- who maintained the electrophysiological apparatus and prepared the riol. 108, 1072-1086. figures. We thank Dr. P. B. Detwiler for reminding us that adenylate 37. Corey, D. P. & Hudspeth, A. J. (1979) Nature (London) 281, kinase never rests. Dr. J. Howard and Mr. G. M. G. Shepherd, as 675-677. well as members of our laboratory group, provided useful comments 38. Ohmori, H. (1985) J. Physiol. (London) 359, 189-217. on the manuscript. This research was supported by National Insti- 39. Hudspeth, A. J. & Jacobs, R. A. (1979) Proc. Natl. Acad. Sci. tutes of Health Grant DC00317. USA 76, 1506-1509. Downloaded by guest on September 23, 2021